gold/stringpool.cc

   1 // stringpool.cc -- a string pool for gold
   2
   3 // Copyright 2006, 2007 Free Software Foundation, Inc.
   4 // Written by Ian Lance Taylor <iant@google.com>.
   5
   6 // This file is part of gold.
   7
   8 // This program is free software; you can redistribute it and/or modify
   9 // it under the terms of the GNU General Public License as published by
  10 // the Free Software Foundation; either version 3 of the License, or
  11 // (at your option) any later version.
  12
  13 // This program is distributed in the hope that it will be useful,
  14 // but WITHOUT ANY WARRANTY; without even the implied warranty of
  15 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16 // GNU General Public License for more details.
  17
  18 // You should have received a copy of the GNU General Public License
  19 // along with this program; if not, write to the Free Software
  20 // Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
  21 // MA 02110-1301, USA.
  22
  23 #include "gold.h"
  24
  25 #include <cstring>
  26 #include <algorithm>
  27 #include <vector>
  28
  29 #include "output.h"
  30 #include "parameters.h"
  31 #include "stringpool.h"
  32
  33 namespace gold
  34 {
  35
  36 template<typename Stringpool_char>
  37 Stringpool_template<Stringpool_char>::Stringpool_template()
  38   : string_set_(), strings_(), strtab_size_(0), next_index_(1),
  39     next_uncopied_key_(-1), zero_null_(true)
  40 {
  41 }
  42
  43 template<typename Stringpool_char>
  44 void
  45 Stringpool_template<Stringpool_char>::clear()
  46 {
  47   for (typename std::list<Stringdata*>::iterator p = this->strings_.begin();
  48        p != this->strings_.end();
  49        ++p)
  50     delete[] reinterpret_cast<char*>(*p);
  51   this->strings_.clear();
  52   this->string_set_.clear();
  53 }
  54
  55 template<typename Stringpool_char>
  56 Stringpool_template<Stringpool_char>::~Stringpool_template()
  57 {
  58   this->clear();
  59 }
  60
  61 // Return the length of a string of arbitrary character type.
  62
  63 template<typename Stringpool_char>
  64 size_t
  65 Stringpool_template<Stringpool_char>::string_length(const Stringpool_char* p)
  66 {
  67   size_t len = 0;
  68   for (; *p != 0; ++p)
  69     ++len;
  70   return len;
  71 }
  72
  73 // Specialize string_length for char.  Maybe we could just use
  74 // std::char_traits<>::length?
  75
  76 template<>
  77 inline size_t
  78 Stringpool_template<char>::string_length(const char* p)
  79 {
  80   return strlen(p);
  81 }
  82
  83 // Compare two strings of arbitrary character type for equality.
  84
  85 template<typename Stringpool_char>
  86 bool
  87 Stringpool_template<Stringpool_char>::string_equal(const Stringpool_char* s1,
  88                                                    const Stringpool_char* s2)
  89 {
  90   while (*s1 != 0)
  91     if (*s1++ != *s2++)
  92       return false;
  93   return *s2 == 0;
  94 }
  95
  96 // Specialize string_equal for char.
  97
  98 template<>
  99 inline bool
 100 Stringpool_template<char>::string_equal(const char* s1, const char* s2)
 101 {
 102   return strcmp(s1, s2) == 0;
 103 }
 104
 105 // Equality comparison function for the hash table.
 106
 107 template<typename Stringpool_char>
 108 inline bool
 109 Stringpool_template<Stringpool_char>::Stringpool_eq::operator()(
 110     const Hashkey& h1,
 111     const Hashkey& h2) const
 112 {
 113   return (h1.hash_code == h2.hash_code
 114           && h1.length == h2.length
 115           && memcmp(h1.string, h2.string,
 116                     h1.length * sizeof(Stringpool_char)) == 0);
 117 }
 118
 119 // Hash function.  The length is in characters, not bytes.
 120
 121 template<typename Stringpool_char>
 122 size_t
 123 Stringpool_template<Stringpool_char>::string_hash(const Stringpool_char* s,
 124                                                   size_t length)
 125 {
 126   // Fowler/Noll/Vo (FNV) hash (type FNV-1a).
 127   const unsigned char* p = reinterpret_cast<const unsigned char*>(s);
 128   if (sizeof(size_t) > 4)
 129     {
 130       size_t result = static_cast<size_t>(14695981039346656037ULL);
 131       for (size_t i = 0; i < length * sizeof(Stringpool_char); ++i)
 132         {
 133           result ^= static_cast<size_t>(*p++);
 134           result *= 1099511628211ULL;
 135         }
 136       return result;
 137     }
 138   else
 139     {
 140       size_t result = 2166136261UL;
 141       for (size_t i = 0; i < length * sizeof(Stringpool_char); ++i)
 142         {
 143           result ^= static_cast<size_t>(*p++);
 144           result *= 16777619UL;
 145         }
 146       return result;
 147     }
 148 }
 149
 150 // Add the string S to the list of canonical strings.  Return a
 151 // pointer to the canonical string.  If PKEY is not NULL, set *PKEY to
 152 // the key.  LENGTH is the length of S in characters.  Note that S may
 153 // not be NUL terminated.
 154
 155 template<typename Stringpool_char>
 156 const Stringpool_char*
 157 Stringpool_template<Stringpool_char>::add_string(const Stringpool_char* s,
 158                                                  size_t len,
 159                                                  Key* pkey)
 160 {
 161   // We are in trouble if we've already computed the string offsets.
 162   gold_assert(this->strtab_size_ == 0);
 163
 164   // The size we allocate for a new Stringdata.
 165   const size_t buffer_size = 1000;
 166   // The amount we multiply the Stringdata index when calculating the
 167   // key.
 168   const size_t key_mult = 1024;
 169   gold_assert(key_mult >= buffer_size);
 170
 171   // Convert len to the number of bytes we need to allocate, including
 172   // the null character.
 173   len = (len + 1) * sizeof(Stringpool_char);
 174
 175   size_t alc;
 176   bool front = true;
 177   if (len > buffer_size)
 178     {
 179       alc = sizeof(Stringdata) + len;
 180       front = false;
 181     }
 182   else if (this->strings_.empty())
 183     alc = sizeof(Stringdata) + buffer_size;
 184   else
 185     {
 186       Stringdata *psd = this->strings_.front();
 187       if (len > psd->alc - psd->len)
 188         alc = sizeof(Stringdata) + buffer_size;
 189       else
 190         {
 191           char* ret = psd->data + psd->len;
 192           memcpy(ret, s, len - sizeof(Stringpool_char));
 193           memset(ret + len - sizeof(Stringpool_char), 0,
 194                  sizeof(Stringpool_char));
 195
 196           if (pkey != NULL)
 197             *pkey = psd->index * key_mult + psd->len;
 198
 199           psd->len += len;
 200
 201           return reinterpret_cast<const Stringpool_char*>(ret);
 202         }
 203     }
 204
 205   Stringdata *psd = reinterpret_cast<Stringdata*>(new char[alc]);
 206   psd->alc = alc - sizeof(Stringdata);
 207   memcpy(psd->data, s, len - sizeof(Stringpool_char));
 208   memset(psd->data + len - sizeof(Stringpool_char), 0,
 209          sizeof(Stringpool_char));
 210   psd->len = len;
 211   psd->index = this->next_index_;
 212   ++this->next_index_;
 213
 214   if (pkey != NULL)
 215     *pkey = psd->index * key_mult;
 216
 217   if (front)
 218     this->strings_.push_front(psd);
 219   else
 220     this->strings_.push_back(psd);
 221
 222   return reinterpret_cast<const Stringpool_char*>(psd->data);
 223 }
 224
 225 // Add a string to a string pool.
 226
 227 template<typename Stringpool_char>
 228 const Stringpool_char*
 229 Stringpool_template<Stringpool_char>::add(const Stringpool_char* s, bool copy,
 230                                           Key* pkey)
 231 {
 232   typedef std::pair<typename String_set_type::iterator, bool> Insert_type;
 233
 234   if (!copy)
 235     {
 236       // When we don't need to copy the string, we can call insert
 237       // directly.
 238
 239       const Key k = this->next_uncopied_key_;
 240       const off_t ozero = 0;
 241       std::pair<Hashkey, Hashval> element(Hashkey(s),
 242                                           std::make_pair(k, ozero));
 243
 244       Insert_type ins = this->string_set_.insert(element);
 245
 246       typename String_set_type::const_iterator p = ins.first;
 247
 248       if (ins.second)
 249         {
 250           // We just added the string.  The key value has now been
 251           // used.
 252           --this->next_uncopied_key_;
 253         }
 254       else
 255         {
 256           gold_assert(k != p->second.first);
 257         }
 258
 259       if (pkey != NULL)
 260         *pkey = p->second.first;
 261       return p->first.string;
 262     }
 263
 264   return this->add_prefix(s, string_length(s), pkey);
 265 }
 266
 267 // Add a prefix of a string to a string pool.
 268
 269 template<typename Stringpool_char>
 270 const Stringpool_char*
 271 Stringpool_template<Stringpool_char>::add_prefix(const Stringpool_char* s,
 272                                                  size_t len,
 273                                                  Key* pkey)
 274 {
 275   // When adding an entry, this will look it up twice in the hash
 276   // table.  The problem is that we can't insert S before we
 277   // canonicalize it by copying it into the canonical list. The hash
 278   // code will only be computed once, so this isn't all that
 279   // expensive.
 280
 281   Hashkey hk(s, len);
 282   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 283   if (p != this->string_set_.end())
 284     {
 285       if (pkey != NULL)
 286         *pkey = p->second.first;
 287       return p->first.string;
 288     }
 289
 290   Key k;
 291   hk.string = this->add_string(s, len, &k);
 292   // The contents of the string stay the same, so we don't need to
 293   // adjust hk.hash_code or hk.length.
 294
 295   const off_t ozero = 0;
 296   std::pair<Hashkey, Hashval> element(hk, std::make_pair(k, ozero));
 297
 298   typedef std::pair<typename String_set_type::iterator, bool> Insert_type;
 299   Insert_type ins = this->string_set_.insert(element);
 300   gold_assert(ins.second);
 301
 302   if (pkey != NULL)
 303     *pkey = k;
 304   return hk.string;
 305 }
 306
 307 template<typename Stringpool_char>
 308 const Stringpool_char*
 309 Stringpool_template<Stringpool_char>::find(const Stringpool_char* s,
 310                                            Key* pkey) const
 311 {
 312   Hashkey hk(s);
 313   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 314   if (p == this->string_set_.end())
 315     return NULL;
 316
 317   if (pkey != NULL)
 318     *pkey = p->second.first;
 319
 320   return p->first.string;
 321 }
 322
 323 // Comparison routine used when sorting into an ELF strtab.  We want
 324 // to sort this so that when one string is a suffix of another, we
 325 // always see the shorter string immediately after the longer string.
 326 // For example, we want to see these strings in this order:
 327 //   abcd
 328 //   cd
 329 //   d
 330 // When strings are not suffixes, we don't care what order they are
 331 // in, but we need to ensure that suffixes wind up next to each other.
 332 // So we do a reversed lexicographic sort on the reversed string.
 333
 334 template<typename Stringpool_char>
 335 bool
 336 Stringpool_template<Stringpool_char>::Stringpool_sort_comparison::operator()(
 337   const Stringpool_sort_info& sort_info1,
 338   const Stringpool_sort_info& sort_info2) const
 339 {
 340   const Hashkey& h1(sort_info1->first);
 341   const Hashkey& h2(sort_info2->first);
 342   const Stringpool_char* s1 = h1.string;
 343   const Stringpool_char* s2 = h2.string;
 344   const size_t len1 = h1.length;
 345   const size_t len2 = h2.length;
 346   const size_t minlen = len1 < len2 ? len1 : len2;
 347   const Stringpool_char* p1 = s1 + len1 - 1;
 348   const Stringpool_char* p2 = s2 + len2 - 1;
 349   for (size_t i = minlen; i > 0; --i, --p1, --p2)
 350     {
 351       if (*p1 != *p2)
 352         return *p1 > *p2;
 353     }
 354   return len1 > len2;
 355 }
 356
 357 // Return whether s1 is a suffix of s2.
 358
 359 template<typename Stringpool_char>
 360 bool
 361 Stringpool_template<Stringpool_char>::is_suffix(const Stringpool_char* s1,
 362                                                 size_t len1,
 363                                                 const Stringpool_char* s2,
 364                                                 size_t len2)
 365 {
 366   if (len1 > len2)
 367     return false;
 368   return memcmp(s1, s2 + len2 - len1, len1 * sizeof(Stringpool_char)) == 0;
 369 }
 370
 371 // Turn the stringpool into an ELF strtab: determine the offsets of
 372 // each string in the table.
 373
 374 template<typename Stringpool_char>
 375 void
 376 Stringpool_template<Stringpool_char>::set_string_offsets()
 377 {
 378   if (this->strtab_size_ != 0)
 379     {
 380       // We've already computed the offsets.
 381       return;
 382     }
 383
 384   const size_t charsize = sizeof(Stringpool_char);
 385
 386   // Offset 0 may be reserved for the empty string.
 387   off_t offset = this->zero_null_ ? charsize : 0;
 388
 389   // Sorting to find suffixes can take over 25% of the total CPU time
 390   // used by the linker.  Since it's merely an optimization to reduce
 391   // the strtab size, and gives a relatively small benefit (it's
 392   // typically rare for a symbol to be a suffix of another), we only
 393   // take the time to sort when the user asks for heavy optimization.
 394   if (parameters->optimization_level() < 2)
 395     {
 396       for (typename String_set_type::iterator curr = this->string_set_.begin();
 397            curr != this->string_set_.end();
 398            curr++)
 399         {
 400           if (this->zero_null_ && curr->first.string[0] == 0)
 401             curr->second.second = 0;
 402           else
 403             {
 404               curr->second.second = offset;
 405               offset += (curr->first.length + 1) * charsize;
 406             }
 407         }
 408     }
 409   else
 410     {
 411       size_t count = this->string_set_.size();
 412
 413       std::vector<Stringpool_sort_info> v;
 414       v.reserve(count);
 415
 416       for (typename String_set_type::iterator p = this->string_set_.begin();
 417            p != this->string_set_.end();
 418            ++p)
 419         v.push_back(Stringpool_sort_info(p));
 420
 421       std::sort(v.begin(), v.end(), Stringpool_sort_comparison());
 422
 423       for (typename std::vector<Stringpool_sort_info>::iterator last = v.end(),
 424              curr = v.begin();
 425            curr != v.end();
 426            last = curr++)
 427         {
 428           if (this->zero_null_ && (*curr)->first.string[0] == 0)
 429             (*curr)->second.second = 0;
 430           else if (last != v.end()
 431                    && is_suffix((*curr)->first.string,
 432                                 (*curr)->first.length,
 433                                 (*last)->first.string,
 434                                 (*last)->first.length))
 435             (*curr)->second.second = ((*last)->second.second
 436                                       + (((*last)->first.length
 437                                           - (*curr)->first.length)
 438                                          * charsize));
 439           else
 440             {
 441               (*curr)->second.second = offset;
 442               offset += ((*curr)->first.length + 1) * charsize;
 443             }
 444         }
 445     }
 446
 447   this->strtab_size_ = offset;
 448 }
 449
 450 // Get the offset of a string in the ELF strtab.  The string must
 451 // exist.
 452
 453 template<typename Stringpool_char>
 454 off_t
 455 Stringpool_template<Stringpool_char>::get_offset(const Stringpool_char* s)
 456   const
 457 {
 458   gold_assert(this->strtab_size_ != 0);
 459   typename String_set_type::const_iterator p = this->string_set_.find(s);
 460   if (p != this->string_set_.end())
 461     return p->second.second;
 462   gold_unreachable();
 463 }
 464
 465 // Write the ELF strtab into the buffer.
 466
 467 template<typename Stringpool_char>
 468 void
 469 Stringpool_template<Stringpool_char>::write_to_buffer(unsigned char* buffer,
 470                                                       size_t bufsize)
 471 {
 472   gold_assert(this->strtab_size_ != 0);
 473    // Quiet the compiler in opt mode.
 474   if (bufsize < static_cast<size_t>(this->strtab_size_))
 475     gold_assert(bufsize >= static_cast<size_t>(this->strtab_size_));
 476   if (this->zero_null_)
 477     buffer[0] = '\0';
 478   for (typename String_set_type::const_iterator p = this->string_set_.begin();
 479        p != this->string_set_.end();
 480        ++p)
 481     {
 482       const int len = (p->first.length + 1) * sizeof(Stringpool_char);
 483       gold_assert(p->second.second + len <= this->strtab_size_);
 484       memcpy(buffer + p->second.second, p->first.string, len);
 485     }
 486 }
 487
 488 // Write the ELF strtab into the output file at the specified offset.
 489
 490 template<typename Stringpool_char>
 491 void
 492 Stringpool_template<Stringpool_char>::write(Output_file* of, off_t offset)
 493 {
 494   gold_assert(this->strtab_size_ != 0);
 495   unsigned char* view = of->get_output_view(offset, this->strtab_size_);
 496   this->write_to_buffer(view, this->strtab_size_);
 497   of->write_output_view(offset, this->strtab_size_, view);
 498 }
 499
 500 // Print statistical information to stderr.  This is used for --stats.
 501
 502 template<typename Stringpool_char>
 503 void
 504 Stringpool_template<Stringpool_char>::print_stats(const char* name) const
 505 {
 506 #if defined(HAVE_TR1_UNORDERED_MAP) || defined(HAVE_EXT_HASH_MAP)
 507   fprintf(stderr, _("%s: %s entries: %zu; buckets: %zu\n"),
 508           program_name, name, this->string_set_.size(),
 509           this->string_set_.bucket_count());
 510 #else
 511   fprintf(stderr, _("%s: %s entries: %zu\n"),
 512           program_name, name, this->table_.size());
 513 #endif
 514   fprintf(stderr, _("%s: %s Stringdata structures: %zu\n"),
 515           program_name, name, this->strings_.size());
 516 }
 517
 518 // Instantiate the templates we need.
 519
 520 template
 521 class Stringpool_template<char>;
 522
 523 template
 524 class Stringpool_template<uint16_t>;
 525
 526 template
 527 class Stringpool_template<uint32_t>;
 528
 529 } // End namespace gold.